ATHI is a framework for discussing AI Threats in a clear and simple way. It's designed to help policymakers , lawmakers make better policy and laws around AI, and also to help the public understand and communicate concenrs in a more universal way.
You use ATHI by filling in the following structure when discussing an AI Threat.
When one places their concerns in this schema it turns the concern into a narrative--a story--that can better be understood by the listener.
Here's an example of using ATHI to discuss a data poisoning attack.
To contribute, simply submit PRs to add categories to the Actors.md, Techniques.md, Harms.md, and Impacts.md files.
This will allow the community to expand the project to become more useful over time.